All types of international economic integration provoke interest because they, both promote and restrict trade at the same time. Trade is liberalized, at least partly, among the participating countries, while it is also distorted with third countries as there are various barriers between the integrated group and the rest of the world. On those grounds the analysis of international economic integration is delicate, complex and speculative. A customs union is the type of integration that has received the most attention in research and is the most rigorously developed branch of the neo-classical theory. This chapter is limited to an analysis of the basics of the static and dynamic models of customs unions.
The tariff system may discriminate between commodities and/or countries. Commodity discrimination takes place when different rates of import duty are charged on different commodities. Country discrimination is found when the same commodity is subject to different rates of duty on the basis of country of origin. Lipsey (1960, p. 496) defined the theory of customs unions as a branch of tariff theory which deals with the effect of geographically discriminatory changes in trade barriers. While this is true in the static sense, however, in a dynamic setting a customs union may be, among other things, a means for economic development.
The efficiency criterion used most often in economics is that of Pareto optimality. An allocation of resources is said to be Pareto-optimal if there does not exist another feasible allocation in which some agents would be better off (in a welfare sense) and no agents worse off. By a judicious definition of welfare, the Pareto-optimal allocation is that allocation which best satisfies social objectives. Pareto optimality is achieved exclusively in the state of free trade and free factor mobility (the first best solution), so that other states, in which there are distortions (tariffs, subsidies, taxes, monopolies, externalities, minimum wages, local content requirements, to mention just a few), are sub-optimal. It may happen that the Pareto-optimal allocation can not be achieved because of one or several distortions. Can a second best position be attained by satisfying the remaining Pareto